You have to remember that species is a human category, and is rather fuzzy. Biologists are forever arguing about it. But in general, behavioral or even geographic isolation is sufficent for speciation. An example: blue-winged and golden winged warblers, which have different but overlapping ranges. In the overlap region, one sees first and second generation hybrids, which are so different from both parents that they were given their own species names before we found out what was going on. The parent species and the hybrids are interfertile, without very much behavioral isolation. Yet we consider them two species within one superspecies.
It's very arbitrary. The northern junco, for example, has well-defined subspecies that are also quite different. Probably because the hybrids tend to vary rather continuously rather than having well-defined plumages, we consider the slate colored junco of the east, the pink-sided junco of the Black Hills, and the Oregon junco of the west all part of the same species. I live in Nebraska, and my yard is full of mutts that can't easily be fit into one of the canonical subspecies.
A large number of American duck species are interfertile. Mallards and Pintails, which are so different any experienced hunter can distinguish them at hundreds of yards, form perfectly fertile hybrids. The American Ruddy Duck has been introduced to Europe and has been hybridizing with the white tailed duck; Spanish game wardens are shooting anything that looks like a ruddy X white-tailed to try to preserve the white-tailed duck gene pool. Black ducks and mallards have fertile hybrids, and currently the black duck population is being overwhelmed by mallard genes.
The point is, if two groups are behaviorally or geographically isolated, it can keep them distinct just as easily as if they could not produce viable offspring. And it would often be almost impossible, in practice, to find out if two species that don't want to mate would give fertile offspring if they did mate. Moreover, it would make no sense to consider the mallard, pintail and wigeon the same species.